diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error sign_error qqplot_mae acf_mae extremogram_mae
nv.cesm2.ssp370 2.77% 0.913 0.663 0.316 0.126 0.039 0.121 0.117
xgboost.cesm2.ssp370 3.23% 0.897 0.595 0.366 0.101 0.027 0.118 0.096
nv.cesm2.ssp585 3.45% 0.919 0.665 0.315 0.125 0.036 0.127 0.124
xgboost.cesm2.ssp585 4.30% 0.914 0.600 0.376 0.102 0.024 0.110 0.103
xgboost.cesm2.ssp245 6.55% 0.869 0.600 0.357 0.101 0.026 0.109 0.092
nv.cesm2.ssp245 6.78% 0.850 0.654 0.302 0.125 0.040 0.130 0.109
nv.mri_esm2_0.ssp370 14.44% 0.808 0.633 0.275 0.124 0.041 0.132 0.106
xgboost.mri_esm2_0.ssp370 15.88% 0.802 0.561 0.363 0.110 0.029 0.099 0.089
nv.ec_earth3.ssp434 -18.17% 0.902 0.695 0.295 0.121 0.062 0.161 0.103
xgboost.ec_earth3.ssp434 -19.53% 0.949 0.670 0.373 0.098 0.045 0.145 0.088
nv.mri_esm2_0.ssp434 26.46% 0.753 0.590 0.276 0.129 0.047 0.128 0.135
xgboost.mri_esm2_0.ssp434 26.77% 0.745 0.507 0.370 0.114 0.041 0.101 0.120
cnn.mri_esm2_0.ssp370 -27.10% 1.002 1.106 0.313 0.124 0.142 0.359 0.101
xgboost.mri_esm2_0.ssp245 28.28% 0.714 0.494 0.371 0.114 0.043 0.120 0.136
nv.mri_esm2_0.ssp245 28.30% 0.728 0.577 0.281 0.130 0.048 0.144 0.161
cnn.mri_esm2_0.ssp434 -29.91% 1.006 1.145 0.323 0.126 0.146 0.341 0.079
cnn.mri_esm2_0.ssp245 -31.18% 1.026 1.155 0.310 0.126 0.143 0.345 0.092
cnn.ec_earth3.ssp434 -31.77% 1.019 1.143 0.326 0.127 0.150 0.342 0.072
cnn.cesm2.ssp585 -34.43% 1.094 1.287 0.334 0.123 0.145 0.275 0.062
cnn.cesm2.ssp245 -42.42% 1.115 1.292 0.326 0.123 0.148 0.288 0.085
cnn.cesm2.ssp370 -42.43% 1.119 1.312 0.320 0.120 0.147 0.268 0.076

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram